The AI Adoption Paradox: Why 97% of Businesses Are Using AI But Customers Are Pushing Back

Nearly every business is racing to put AI between itself and its customers — but the customers themselves are asking companies to stop. That tension is quietly reshaping how service businesses win repeat work.

According to Sinch's survey of 1,600 business leaders, 97% of businesses plan to use AI in customer communications in 2025. Yet Gartner research found that 64% of customers would prefer companies didn't use AI for customer service at all. Meanwhile, 73% of support professionals flag customer resistance to AI as a major obstacle.

The result is a genuine paradox: adoption is near-universal while trust is eroding. As one analysis of leading AI deployments puts it, the gap between promise and perception comes down to implementation. AI that deflects volume without human judgment tends to alienate the very customers it's meant to serve.

The stakes are highest in permission-sensitive industries — home services, dental and med spa clinics, healthcare — where outreach touches real relationships and regulated data. Compliance adds another layer: since January 2024, AI-generated voices are regulated as "artificial voices" under the TCPA, with penalties up to $1,500 per call or text for willful violations, and clear, documented consumer consent remains the cornerstone of compliance.

What resolves the paradox isn't less technology — it's the right division of labor:

• Automation handles the scale: list segmentation, timing, and multi-channel campaign delivery
• Humans handle the judgment: what to say, to whom, and when a conversation needs a real person
• The customer controls the channel: 58% of consumers want to choose how they're contacted

That's the model behind the companies getting AI right. aXcelerate, for instance, rejected pure deflection tools because they didn't align with its human-first support philosophy, and 75% of CX leaders now see AI as a force for amplifying human intelligence rather than replacing it.

For service businesses, the economics make the trust question even more urgent. The most common reason customers leave is perceived indifference — not bad service or price — and acquisition runs 5-25x the cost of retention. A lapsed customer won by a thoughtful, human-approved outreach is worth far more than one lost to a robotic script.

This is why CallMyCustomers insists the owner approves every script, offer, and message before anything is sent. Real humans, real judgment — automation handles the scale, people handle the judgment. In a market where 97% adoption meets 64% distrust, the businesses that keep humans in the loop are the ones that keep customers for the long run.

Turn past customers, old quotes, and inactive members into booked work — approved by you, run by us. Get your free list review and see what your customer list can produce before you spend a dollar.

What the Top 10 AI-Using Companies Get Right: Augmentation, Not Replacement

Many companies chase AI’s promise of scale, but the most successful deployments recognize that automation works best when it augments, not replaces, human judgment. The leading AI-using companies in customer service have learned to blend technology with human oversight, using AI to handle repetitive tasks while preserving the empathy and decision-making that customers value. This approach directly addresses a growing trust gap: while 97% of businesses plan to use AI in customer communications by 2025, 64% of customers would prefer companies didn’t use AI for service at all. The winners aren’t those removing humans from the loop — they’re the ones using AI to make human agents more effective.

Assembled’s case studies reveal a clear pattern: top performers use AI to solve specific operational challenges like workforce management, real-time forecasting, and escalation reduction, not just to deflect volume. Intercom uses its Fin AI agent for simple queries, freeing up teams to focus on complex issues and saving over 5 hours per week in scheduling time. Brooks leverages AI for real-time scheduling and forecasting, cutting average phone wait times by 66%. aXcelerate, which initially considered deflection tools but rejected them to preserve its human-first philosophy, uses Assembled AI to reduce new-agent training time by 50%. These companies aren’t chasing AI for its own sake — they’re applying it where it removes friction, improves accuracy, and scales capacity without sacrificing service quality.

The results speak for themselves when AI supports, rather than supplants, human agents. Thrasio deployed Assembled AI across 190 brands handling 70,000+ monthly interactions, lifting CSAT from 87% to 97%, cutting response time to 1 minute, and resolution time to 12 minutes — all while saving $1.8 million annually. Honeylove tailored its AI to reduce escalations, achieving a 20% drop in escalated cases and boosting power users’ solves per hour by 54% in just five months. These gains aren’t theoretical; they come from AI handling routine inquiries, predicting demand, and surfacing insights — all while keeping humans in charge of nuanced conversations, judgment calls, and relationship-building.

Contrast this with AI-only models that attempt full automation without human guardrails. Such approaches often increase frustration, especially when customers sense indifference or receive generic, tone-deaf responses. As Ainora notes, the most common reason customers leave isn’t bad service or high prices — it’s perceived indifference. AI that can’t adapt to context or escalate appropriately risks reinforcing that feeling. The companies seeing real success use AI to reduce the burden on teams — lowering handle times, improving first-response rates, and cutting training ramp-up — so human agents can focus on what they do best: listening, empathizing, and solving problems that require judgment.

For businesses like those served by CallMyCustomers — home services, clinics, salons, and other repeat-revenue providers — this balance is especially critical. Reactivation campaigns succeed not because they blast messages at scale, but because they feel useful, not pushy. Every script, offer, and message is approved by the owner, ensuring tone and timing align with the business’s voice. Automation handles the outreach volume; humans handle the judgment. This mirrors the winning pattern seen in Assembled’s top 10: AI as an enabler of human excellence, not a replacement for it. When scale and sensitivity work together, the result isn’t just efficiency — it’s trust, loyalty, and measurable gains in CSAT, response time, and cost savings.

Why Reactivation Is the Highest-ROI Use Case for AI in Customer Communications

Most businesses spend the overwhelming majority of their marketing budget chasing strangers while a list of people who already trust them sits dormant. That imbalance is exactly why reactivation has emerged as one of the highest-ROI applications of AI in customer communications — and why the numbers behind it deserve a closer look.

The economics start with a well-documented cost gap. Harvard Business Review research indicates that acquiring a new customer costs 5 to 25 times more than retaining or reactivating an existing one. Meanwhile, 73% of marketers report that customer acquisition costs keep rising, making the dormant customer list one of the few undervalued assets left in most businesses.

Channel choice matters just as much as budget. Research on AI-driven win-back campaigns shows that phone calls convert 8–12x better than emails for reactivation, while generic "we miss you" emails recover only 3–8% of lapsed customers. Add it up, and AI-powered reactivation campaigns can deliver ROI of up to 800% compared with new-customer acquisition — largely because the audience already knows and once chose your business.

The economics that make reactivation attractive:

• Reactivating a customer is roughly 5x cheaper than acquiring a new one, with HBR pegging the gap as high as 25x
• Businesses running automated reactivation report an average 20% revenue uplift from reengaged customers and up to 30% savings on marketing spend
• Personalized outreach can lift customer engagement by as much as 20%, per McKinsey

The most common reason customers go quiet isn't bad service or price — it's perceived indifference. AI is well suited to fixing that at scale: segmenting a list by recency, flagging old quotes and expiring memberships, and timing outreach so a message lands when it feels useful rather than pushy.

This is where the human-plus-AI pattern seen among top-deploying companies matters. Just as aXcelerate rejected pure deflection tools in favor of a human-first approach, CallMyCustomers applies AI to the mechanical work — list segmentation, campaign timing, multi-channel scale — while the owner approves every script and offer before anything goes out. The judgment stays human; the automation handles the volume.

That balance also matters for trust. With 64% of customers saying they'd prefer companies didn't use AI for customer service at all, a model where real people make the calls and the business signs off on every message is the difference between reactivation that feels like a courtesy and outreach that feels like a robot.

The result is a second revenue engine that runs alongside acquisition — turning past customers, old quotes, and inactive members into booked work, with the economics firmly on your side.

Compliance and Consent as Non-Negotiables in AI-Powered Outreach

AI can send thousands of messages in minutes — but a single non-compliant call can cost more than the entire campaign earns. As companies race to automate outreach, the legal fine print is where careless operators get burned.

The regulatory stakes are real. Since January 2024, the TCPA classifies AI-generated voice as an "artificial voice," bringing it squarely under federal telemarketing rules, with penalties reaching $1,500 per call or text for willful violations, according to TCPA compliance guidance. The same guidance is blunt about what matters most: clear, documented consumer consent remains the cornerstone of compliance.

The landscape is getting harder to navigate, not easier. The Supreme Court's June 2025 McLaughlin v. McKesson ruling created a fragmented compliance environment where FCC interpretations are no longer automatically binding on federal courts, meaning businesses can't simply assume one regulator's playbook protects them. Add in the trust problem — Gartner research found 64% of customers would prefer companies didn't use AI in customer service at all — and it's clear that compliance is no longer just legal hygiene. It's a competitive differentiator.

This is where provider selection gets serious. When 97% of businesses plan to use AI in customer communications, the question isn't whether to automate — it's whether your provider treats permission as a feature or an afterthought. Ask these questions before signing anything:

• Does the provider work only from verified customer lists, or will it call anyone with a pulse?
• Are opt-outs honored immediately, with documentation?
• Who approves each script and offer before it goes out — you, or an algorithm?
• For clinical clients, does outreach operate under the required privacy agreements (BAA/HIPAA)?

CallMyCustomers was built around these questions rather than retrofitting answers to them. The service works exclusively from lists of real customers, honors opt-outs the moment they arrive, and requires the owner to sign off on every script, offer, and message before anything sends — "We plan the campaign together, you sign off, we run it." For dental, med spa, and clinic clients, patient outreach operates under the required clinical privacy standards, and the booking flow collects explicit consent.

Done right, compliance becomes a trust signal your customers can feel. A permissioned call to someone who already knows your business feels useful, not intrusive — and it keeps you on the right side of a regulatory line that grows sharper every year. Want to see what your list can produce within those guardrails? Get your free list review and find out what your past customers are worth before you spend a dollar.